In his recent contribution to the Center for Progressive Reform’s symposium on the Milward case, Professor Cranor suggests that the International Agency for Research on Cancer (IARC) uses weight of the evidence (WOE) in its carcinogenicity determinations. See Carl F. Cranor, “Milward v. Acuity Specialty Products: Advances in General Causation Testimony in Toxic Tort Litigation,” PDF 3 Wake Forest J. L. & Policy 105 (2013)[hereinafter cited as Cranor] Cranor’s suggestion is demonstrably wrong.
The IARC process is described in several places, but the definitive presentation of the IARC’s goals and methods is set out in a document known as the “Preamble.” World Health Organization, IARC Monographs on the Evaluation of Carcinogenic Risks to Humans — Preamble (2006) [cited herein as Preamble]. There is no mention of WOE in the Preamble.
The IARC process consists upon assessments of carcinogenicity of substances or exposure circumstances, and categorization into specified groups:
| IARC Category |
Verbal Description |
IARC “Findings” |
| Group 1 | [Known] Carcinogenic to humans |
111 |
| Group 2A | Probably carcinogenic to humans |
65 |
| Group 2B | Possibly carcinogenic to humans |
274 |
| Group 3 | Not classifiable as to its carcinogenicity to humans |
504 |
| Group 4 | Probably not carcinogenic to humans |
1 |
The IARC operative definitions that a substance to a category are highly stylized and unique to IARC. The definitions do not coincide with ordinary language definitions or general scientific usage. Only one substance is categorized as “probably not carcinogenic to humans” is Caprolactam.
Alas, oxygen, nitrogen, carbon dioxide, sugar, table salt, water, and many other exposures we all experience, and even require, do not make it to “probably not carcinogenic.” This fact should clue in the casual reader that the IARC classifications are greatly influenced by the precautionary principle. There is nothing wrong with this influence, as long as we realize that IARC categorizations do not necessarily line up with scientific determinations.
Cranor attempts to exploit IARC classifications and their verbiage, but in doing so he misrepresents the IARC enterprise. For instance, his paper for the CPR symposium strongly suggests that a case involving a Group 2A carcinogen would necessarily satisfy the preponderance of evidence standard common in civil cases because the IARC denominates the substance or exposure circumstance as “probably carcinogenic to humans.” This suggestion is wrong because of the technical, non-ordinary language meanings given to “probably” and “known.” The IARC terminology involves a good bit of epistemic inflation. Consider first what it means for something to be “probably” a carcinogen:
“Group 2.
This category includes agents for which, at one extreme, the degree of evidence of carcinogenicity in humans is almost sufficient, as well as those for which, at the other extreme, there are no human data but for which there is evidence of carcinogenicity in experimental animals. Agents are assigned to either Group 2A (probably carcinogenic to humans) or Group 2B (possibly carcinogenic to humans) on the basis of epidemiological and experimental evidence of carcinogenicity and mechanistic and other relevant data. The terms probably carcinogenic and possibly carcinogenic have no quantitative significance and are used simply as descriptors of different levels of evidence of human carcinogenicity, with probably carcinogenic signifying a higher level of evidence than possibly carcinogenic.”
Preamble at 22, § 6(d). So probably does not mean “more likely than not,” and “possibly” means something even less than some unspecified level of probability. An IARC classification of 2A will not help a plaintiff reach the jury because it does not connote more likely than not.
A category I finding is usually described as a “known” carcinogen, but the reality is that there may still be a good deal of epistemic uncertainty over the classification:
“Group 1: The agent is carcinogenic to humans.
This category is used when there is sufficient evidence of carcinogenicity in humans. Exceptionally, an agent may be placed in this category when evidence of carcinogenicity in humans is less than sufficient but there is sufficient evidence of carcinogenicity in experimental animals and strong evidence in exposed humans that the agent acts through a relevant mechanism of carcinogenicity.”
Preamble at 22, § 6(d).
Again, the precautionary nature of the categorization should be obvious. Knowledge of carcinogenicity is equated to sufficient evidence, which leaves open whether there is a body of contradictory evidence. The IARC’s definition of “sufficiency” does place some limits on what may affirmatively count as “sufficient” evidence:
“Sufficient evidence of carcinogenicity: The Working Group considers that a causal relationship has been established between exposure to the agent and human cancer. That is, a positive relationship has been observed between the exposure and cancer in studies in which chance, bias and confounding could be ruled out with reasonable confidence. A statement that there is sufficient evidence is followed by a separate sentence that identifies the target organ(s) or tissue(s) where an increased risk of cancer was observed in humans. Identification of a specific target organ or tissue does not preclude the possibility that the agent may cause cancer at other sites.”
Preamble (2006), at 19, § 6(a). This definition hardly helps Cranor in his attempt to defend bad science. Scientists may reasonably disagree over what is sufficient evidence, but the IARC requires, at a minimum, that “chance, bias and confounding could be ruled out with reasonable confidence.” Id. Ruling out chance, of course, introduces considerations of statistical significance, multiple comparisons, and the like. Ruling out bias and confounding with confidence is an essential part of the IARC categorization process, just as it is an essential part of the scientific process. Reviewing the relied upon studies for whether they ruled out chance, bias, and confounding, was precisely what the Supreme Court did in General Electric v. Joiner, and what the current statute, Federal Rule of Evidence 702, now requires. Failing to review the extant epidemiologic studies for their ability to rule out chance, bias, and confounding is exactly what the district court judge did in Milward.
IARC and Conflicts of Interest – Nemo iudex in causa sua
Holding out the IARC process as exemplifying scientific method involves other controversial aspects of the process. IARC’s classifications are determined by “working groups” that review the available scientific literature on an agent’s carcinogenicity. Members of these of groups are selected in part because they have “have published significant research related to the carcinogenicity of the agents being reviewed… .” Preamble at 5. See also Vincent Cogliano, Robert A. Baan, Kurt Straif, et al., “The science and practice of carcinogen identification and evaluation,” 112 Envt’l Health Persp. 1269, 1273 (2004).
While the IARC tries hard to avoid apparent financial conflicts of interest, its approach to selecting voting members of the working groups invites a more pervasive, more corrupting influence: working group members must vote on the validity of their own research. The prestige of their own research will thus be directly affected by the group’s vote, as well as by the analysis in the resulting IARC monograph. Many writers have criticized this approach. See, e.g., Paolo Boffetta, Joseph McLaughlin, Carlo La Vecchia, Robert Tarone, Loren Lipworth, and William Blot, “A further plea for adherence to the principles underlying science in general and the epidemiologic enterprise in particular,” 38 Internat’l J. Epidemiol. 678 (2009); Michael Hauptmann & Cecile Ronckers, “A further plea for adherence to the principles underlying science in general and the epidemiologic enterprise in particular,” 39 Internat’l J. Epidemiol. 1677 (2010).
Notably absent from Cranor’s defense of using bad science and incomplete evidence is his disregard of systematic reviews and meta-analysis. Although “agency” science is a weak shadow of the real thing, even federal agencies have come to see the importance of using principles of systematic reviews in their assessments of science for policy purposes. See, e.g., FDA, Guidance for industry evidence-based review system for the scientific evaluation of health claims (2009). Currently underway at the National Toxicology Program’s Office of Health Assessment and Translation (OHAT) is an effort to implement systematic review methodology in the Program’s assessments of potential human health hazards. That the NTP is only now articulating an OHAT Evaluation Process, incorporating principles of systematic review, suggests that something less rigorous has been used previously. See Federal Register Notice , 78 Fed. Reg. 37 (Feb. 25, 2013).
No one should be fooled by Cranor’s attempt to pass off precautionary judgments as scientific determinations of causality.
